Response feature analyses for repeated measures behavioral data

Repeated measures designs are a common experimental setup in biology, from ecology to biomedical sciences. Traditional statistical techniques for analyzing such data, such as repeated measures ANOVA (RMANOVA), have limitations that prevent them from being used to answer many questions relevant to measurements taken over time. Response feature analysis (RFA) techniques can provide an ideal means for answering those questions by analyzing data based on a creatively chosen summary function of the individual subject's data. The process and rationale for the RFA is laid out in a step-by-step manner and is demonstrated on an example of behavioral data from a rodent spinal cord injury study. In the original analysis, a difference in average functional score was only found at one timepoint early in the study. Using RFA, we show that the female rats reached 90% of their peak recovery 15.3 days (95% CI 1.73, 28.86 days) earlier than the male rats. RFA enables answering the important question 'do the two sexes differ in the speed at which they recover functional ability?', and, by way of demonstration, illustrates its flexibility in answering questions that traditional repeated measures analyses cannot.